ES2604133T3 - Device and method to generate an environment signal - Google Patents

Abstract

A device for generating an ambient signal suitable for output through loudspeakers (Ls, Rs) for which there is no suitable loudspeaker signal, comprising: a transient detector (11) for detecting a transient (22) in a test signal block to obtain a transient period (20) comprising the test signal block; a synthesis signal generator (12) for generating a synthesis signal during the transient period (20), the synthesis signal generator (12) being implemented to generate a synthesis signal comprising a flatter time course than the synthesis signal test in the transient period (20) and whose intensity deviates from an intensity of a part of the test signal before the transient period (20) or after the transient period (20) by less than a predetermined threshold; and a signal substitute (14) to substitute the test signal in the transient period for the synthesis signal to obtain the ambient signal, in which the transient detector (11) is implemented to calculate the weighted high frequency contents for the test signal block (71), in which the weighted high-frequency contents are a weighted sum of the absolute values of all frequency lines in the block with increasing weighting factors from higher frequencies. low to the highest; wherein the transient detector (11) is implemented to compare (73) the weighted high-frequency contents of the block for a floating average value (72) with the weighted high-frequency contents of a plurality of preceding or succeeding blocks of the block without any transient, wherein the transient detector (11) is implemented to detect a transient for the block when the weighted high-frequency contents of the block exceed the floating mean value by more than a threshold (c), wherein the synthesis signal generator is implemented to calculate (74), when the synthesis signal is generated for all spectral values of a short-term spectrum of a plurality of blocks, which form the part of the test signal before the transient period (20) or after the transient period (20), an average value that uses the corresponding spectral values of the plurality of blocks to obtain an average value spectrum for the block, calculate, for the spectral values of the mean value spectrum for the block, the deviations that differ from the spectral values of the mean value spectrum for the block and that are smaller than a maximum deviation (Δmax), and add (75) the deviations and the mean value of the spectral values to obtain the spectral values for the block, and converting (43, 76) the spectral values for the block into a time representation representing the synthesis signal (44).

Description

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

55

60

Device and method to generate an environment signal DESCRIPTION

The present invention relates to the processing of an audio signal and, in particular, to the concepts of generating ambient speaker signals in a multichannel scenario for which no special speaker signal has been transmitted.

Multichannel audio material is increasing in popularity. This has resulted in many end users who now own multi-channel playback systems. This can be attributed mainly to the fact that DVDs are increasing in popularity and that many DVD users are now in possession of 5.1 multi-channel equipment. Reproduction systems of this type, in general, include three speakers L (left), C (center) and R (right), which are normally arranged in front of the user, and two speakers Ls and Rs arranged behind the user, and by usually an LFE channel, which is also known as a low frequency effects channel or subwoofer. A channel scenario of this type is indicated in Figures 10 and 11. While the placement of the loudspeakers L, C, R, Ls, Rs with respect to the user is performed as indicated in Figure 10 and Figure 11 with In order for the user to receive the best possible auditory impression, the placement of the LFE channel (not shown in Figures 10 and 11) is not as important as the finger cannot locate at such low frequencies and the LFE channel can in this way organize in any place where it has no disturbing effect due to its considerable size.

Such a multi-channel system produces several advantages compared to a typical stereo reproduction which is a two-channel reproduction, as shown by way of example in Figure 9.

Outside the optimal central listening position, the result will also be the improved stability of the front hearing impression which is also known as the "front image", due to the central channel. Therefore, the result is greater than a “sweet spot”, the “sweet spot” representing the optimal listening position.

In addition, due to the two rear speakers Ls and Rs, the listener has an improved feeling of "deepening" the audio scene.

However, there is a large amount of audio material in the possession of users or generally available that is only present as stereo material, which therefore has only two channels, namely the left channel and the right channel. Typical sound carriers for stereo parts of this type are compact discs.

In order to reproduce such stereo material through a 5.1 multi-channel audio device, there are two options recommended in accordance with the ITU.

The first option is to play the left and right channels through the left and right speakers of the multi-channel playback system. However, this solution has the disadvantage that the plurality of speakers already present do not make use of them, that is to say, that use is not made in an advantageous manner of the center speaker and two rear speakers are present.

Another option is to convert the two channels to form a multichannel signal. This can take place during playback or by special preprocessing, which makes advantageous use of the six speakers of the 5.1 playback system already present by way of example and therefore results in an improved auditory impression when an up mix is performed. from two channels to five and / or six channels without errors.

Only then will the second option, that is, using all the speakers of the multichannel system, be advantageous compared to the first solution, in the event that no errors occur when performing the upmix. Upward mixing errors of this type can be dramatically worrying when the signals from the rear speakers, which are also known as ambient signals, are not generated in an error-free manner.

One way to perform this so-called upmixing process is known under the key expression "direct environment concept". The direct sound sources are reproduced by the three front channels present in such a way that they are perceived by the user in the same position as in the original two-channel version. The original two-channel version is schematically illustrated in Figure 9 using the example of the different percussion instruments.

Figure 10 shows an up mix version of the concept in which all the original sound sources, that is, the percussion instruments, are played again by the three front speakers L, C and R, in which also the signals Special ambient are emitted by the two rear speakers. Therefore, the

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

55

60

The expression "directed sound source" is used to describe a tone originating solely and directly from the discrete sound source, such as, for example, a percussion instrument or other instrument, or, in general, a special audio object, such as exemplified schematically in figure 9 using a percussion instrument. Any additional sound, such as, for example, due to wall reflections, etc., is not present in said direct sound source. In this scenario, the sound signals emitted by the two rear speakers Ls, Rs in Figure 10 include, or not, only the ambient signals present in the original recording. Ambient signals of this type do not belong to a single source of sound, but contribute to the reproduction of the acoustics of the room of a recording and therefore give rise to the so-called sensation of "productivity in" in the listener.

Another alternative concept called "in the band" concept is schematically illustrated in Figure 11. All types of sound, that is, direct sound sources and ambient-type tones, are placed around the listener. The position of a tone is independent of its characteristic (direct sound sources or ambient-type tones) and only depends on the specific design of the algorithm, as illustrated by way of example in Figure 11. Therefore, it has been determined in Figure 11 by the upmix algorithm that the two instruments 1100 and 1102 are placed laterally in relation to the listener, while the two instruments 1104 and 1106 are placed in front of the user. The result of this is that the two rear speakers Ls, Rs also contain parts of the two instruments 1100 and 1102 and no longer just the ambient type tones, as was the case in Figure 10 where all the same instruments were placed in front of the user.

The specialized publication of C. Avendano and J.M. Jot: “Ambience Extraction and Synthesis from Stereo Signals for Multichannel Audio Mixup”, International IEEE Conference on Acoustics, Speech and Signal Processing, ICASSP 02, Orlando, Fl, May 2002 reveals a frequency domain technology to identify and extract information of environment in stereo audio signals. This concept is based on calculating a coherence between the channels and a non-linear mapping function that is to allow the determination of the time-frequency regions in the stereo signals that mainly include ambient components. Next, the ambient signals are synthesized and used to store the rear channels or "surround sound" channels Ls, Rs (Figures 10 and 11) of a multi-channel playback system.

In the specialized publication of R. Irwan and Ronald M Aarts: “A method to covert stereo to multi-channel sound”, Procedures of the 19th AES International Conference, Schloss Elmau, Germany, June 21-24, pages 139143, 2001, A method is presented to convert a stereo signal into a multichannel signal. The signal for the surround channels is calculated using a cross correlation technique. Principal component analysis (PCA) is used to calculate a vector that indicates a direction of the dominant signal. This vector is then mapped from a representation of two channels to a representation of three channels to produce the three front channels.

The specialized publication of G. Soulodre, “Ambience-Based Up-mixing”, Workshop of “Spatial Coding of Surround Sound: A Progress Report”, 117th AES Convention, San Francisco, CA, US, 2004 unveils a system that produces a multichannel signal from a stereo signal. The signal is divided into so-called source flows and individual environment flows. Based on these flows, a so-called "aesthetic processor" synthesizes the multichannel output signal.

All known technologies try to extract the environment signals from the original stereo signal in different ways or even synthesize them from noise and / or additional information, in which information that is not in the stereo signal can also be used to synthesize the ambient signals. In the end, however, everything is around the extraction of information from the stereo signal and / or the feeding of information to a reproduction scenario, the information not being explicitly present, since usually only one signal is available two-channel stereo and, perhaps, additional information and / or meta information.

From this point of view, the extraction or partial extraction and partial synthesis of such ambient signals is a risky matter, since a user perceives it as annoying if the information of the sound sources is contained in the ambient channels, that the user identifies that it comes directly from the front, that is, from the left channel, the central channel and the right channel. For this reason, a production of the ambient signals becomes very "defensive" in order to ensure that there are no disturbances perceived by the user that are of concern. The other extreme case, when acting too defensively producing the environment signals is an environment signal that is too weak or almost imperceptible to be extracted or the environment signal only comprises the noise, but no more special information so that the Ambient signal contributes very slightly to a pleasure of hearing and in this case it can really be omitted completely.

It is problematic when the environmental signal is produced that, on the one hand, an environmental signal is produced that includes information that goes beyond normal noise, but that the environmental signal does not give rise to audible disturbances, that is, that it should maintain an appropriate measure between audibility and information content.

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

55

60

US 2006/0018486 A1 discloses an efficient filter to generate an artificial environment signal. The filter includes a transient reduction module adapted to reduce the transients in an audio signal, the audio signal comprising one or more channels. In addition, the filter includes a reverberation filter adapted to receive the transiently reduced signal and to generate a reverberation signal comprising one or more channels and corresponding to the transiently reduced signal.

US Patent 4,076,969 A discloses a pulse noise reduction system for detecting sound signals from sparks and unwanted clicks generated by mechanical defects in a recording medium. A pair of correlated signals are generated totally or partially from the recorded sound information by means of a stereo recording device. The output signals of the stereo recording device are combined in such a way that the desired sound signals generated by a horizontal movement on the recording medium are subtracted and the unwanted signals generated by a vertical movement on the recording medium are added. The resulting unwanted signal is detected as a noise signal. A control signal that starts from the leading edge of the unwanted sound signal detected and having a width that is larger than the width of the unwanted sound signal detected is used in the system to eliminate unwanted signals.

The specialized publication "Restauration of Historical Recordings by Means of Digital Signal Processing", preprinted No. 2091, WA Deutsch, et al., 75th Convention of 1984, refers to improving the overall quality of recordings that have deteriorated Due to various disturbances. In particular, impulsive noises generated by scratches and jumps in the recording material must be eliminated. In addition, distortions in the frequency range generated by the recording device must be compensated. In addition, continuous broadband surface noise should be reduced.

The specialized publication "Detection and modeling of fast attack transients", X. Rodet et al., Proc. Of Intl. Computer Music Conf. (ICMC), September 1, 2001, pages 1-4, reveals the detection and modeling of transients based on rapid attacks. Therefore, the energy in a frequency band is examined and a peak is sought that has a triangular shape that is between the previous and subsequent plateaus.

The specialized publication "Enhancement of Audio Signals Using Transient Detection and Modification", AES Convention Document, New York, vol. 6255, 2004, pages 2-11, discloses a processing approach that allows a perceptually pleasant modification of the audio signals through highlighting and suppressing, respectively, the transients. The frequency range analysis that results in a spectral flow parameter is used. This parameter is compared with a threshold to obtain a binary transient detection function.

It is the object of the present invention to provide a concept for producing an environment signal in which audible disturbances are reduced.

This object is achieved by means of a device for producing an ambient signal according to claim 1, a method for producing an ambient signal according to claim 9 or a computer program according to claim 10.

The present invention is based on the finding that the disturbances that are perceived by the listeners as being more negative in the ambient signals being disturbances that give rise to the listener believe that there is a direct sound source in the rear speaker, although I perceive this sound source as coming from the front. The characteristics to perceive direct sound sources are transient processes, that is, signals of fine structures in the time signal related to a (rapid) change along a threshold of alteration from a weak state to a high state or from a high state to a weak state and / or relating to a (strong) increase in energy along a threshold of alteration in the special bands and, in particular, in the upper bands within a given time.

Transient processes of this type are, for example, a starting instrument or a percussion instrument that is struck or the end of a tone that does not fade slowly but stops abruptly. A listener will perceive such transient processes as characteristic of direct sound sources which, according to the invention, are removed from an ambient signal such that the ambient speakers are provided with an ambient signal produced according to the invention that includes transients or only strongly attenuated transients.

According to the invention, it is guaranteed that the suppression of a transient in the ambient signal does not result in a modulation of amplitude too large. It has been discovered according to the invention that variations in amplitude, that is, in sound intensity, although it is not transient, that is, below the transient threshold, but above a certain threshold of variation, they are recognized by the user as being

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

55

60

disturbing and recognized by the listener as disturbances or errors when such amplitude variations result from a simple elimination of a transient in an ambient signal.

According to the invention, in a test signal, a transitional period is detected in which a transitional region is present in the test signal. Subsequently, using a synthesis signal generator, a synthesis signal is produced during the transitional period, the generator being implemented to generate the synthesis signal in such a way that it has a flatter time course than the test signal in the transitional region, also implementing the synthesis signal generator to generate the synthesis signal in such a way that it differs, with respect to the intensity of an anterior or posterior part of the test signal, less than a predetermined threshold. This produced synthesis signal is then used by a signal substitute instead of the test signal in the transitional period to obtain the ambient signal.

Therefore, the extraction of an ambient signal type signal from a two-channel stereo input signal according to the invention is improved, or a post-processing of an existing signal is performed which, for example, is already a raw environment signal extrafda. In the first case, the test signal is the real two-channel stereo signal and / or a respective channel of the two-channel signal, while in the second case the test signal is an extra-ambient signal or a signal of pre-synthesized environment. Therefore, the concept of the invention is specifically useful for the concept of upward mixing that has also been illustrated as "concept of direct environment". The concept of the invention can also be advantageous for the concept of "in the band", since, in this case, it will also give rise to an improved environment signal that, on the one hand, has no more disturbing disturbances but, on the other On the contrary, it still includes enough information in order for a user to benefit from the environment signal.

The generation of the ambient signal of the invention has the result that the ambient signal has no relevant parts of the direct sound sources, in which, in particular, there are no transient contents and / or transients only included in one form. very strongly attenuated. Otherwise, the listener perceives direct sound sources behind himself, which is in conflict with the user experience that normally only perceives sound sources from the front.

In addition, the concept of the invention guarantees that the ambient signal is a continuous uninterrupted continuous tone signal since an interrupted ambient type tone that is simply obtained, for example, when the transients are completely eliminated, is perceived by the user as which is unpleasant or even as an error in the upward mixing process.

In a preferred embodiment of the present invention, an ambient type signal is extracted for the rear channels from the stereo signal to achieve a direct ambient upward mixing process. In order to achieve this, only the uncorrelated signal components are used as an example or, as a simple solution, the difference between the original right and left channels is simply used. If the rear channels are produced in this way, they will often comprise transient type components of the direct sound sources. These transients may be tones, such as, for example, beginnings of notes or parts of percussion instruments. A transient perceived as being behind the listener, while a direct sound source (to which the transient normally belongs) is placed in front of the listener, has a negative impact on the location of the direct sound source. Therefore, the direct sound source seems to be either wider than the original or is perceived, which is even more harmful, as an independent direct sound source behind the user, in which both effects are very unfavorable, in particular, for the concept of direct environment.

According to the invention, these problems are addressed by suppressing the transients in the ambient type signal and minimizing the effect of this suppression on the remaining signal, that is, maintaining the continuity of the signal, allowing only limited intensity variations for the transitional period

In the preferred embodiment of the present invention, the signal produced during the transitional period is mixed, before being used by the signal substitute, with the signal originally present in the transitional period, which is what is achieved, for example, by a overlay processing. Alternatively or additionally, cross-attenuation can be performed to suppress or at least reduce discontinuities at the edges of the transitional period, in order to perform a slowly cross-attenuation in a cross-attenuation region from the signal before the transitional period to the signal in the transitional period or to attenuate it again slowly from the transitional period.

In particular, it is preferable to attenuate from the transitional period to the original signal when no more transient ones are detected for a disturbance-free auditory impression, since it is necessary to ensure that no sizzling or similar effect is produced by the transition from the signal Synthesis to the original test signal when there is a test signal not vitiated by disturbances.

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

55

60

In other preferred embodiments of the present invention, the manipulation of the signal in the transitional period in the frequency domain is performed by randomization of signals of spectral values or, more generally, phases of the spectral values, which inevitably results to soften the temporal fine structure of this signal manipulated in the frequency domain. The additional spectral processing makes a prediction as to the frequency of the spectral values and then uses the prediction spectral values as spectral values of the synthesis signal, since the prediction regarding the frequency results results in smoothing the corresponding time signal.

In order to suppress the transients while maintaining or only slightly influencing them, it is preferred to change the intensity of the transitional period by a maximum of +/- 50%, that is, by limiting the variation of the spectral values from one block to the next , in which this limitation can take place globally, that is, the same for all spectral values or selectively, that is, only for certain spectral values that specifically comprise a large variation.

Preferred embodiments of the present invention will be detailed below with reference to the accompanying drawings in which:

figure 1

figure 2a

figure 2b figure 3

figure 4

Figure 5a Figure 5b Figure 5c Figure 6 Figure 7

figure 8

figure 9 figure 10

figure 11

it is a block circuit diagram of the device of the invention for producing an ambient signal;

it is a schematic illustration of the block process with non-overlapping blocks, but with a cross-attenuation region;

it is a schematic illustration of the generation of synthesis signal with superimposed blocks;

shows a special cross-attenuation implementation with a slow input function and a slow shutdown function that can be used for Figure 2a or Figure 2b;

it is a block circuit diagram of a preferred embodiment that includes processing in the frequency domain;

shows an alternative implementation of frequency domain processing;

shows other alternative frequency domain processing;

shows a preferred implementation of intensity-based processing;

shows an implementation to maintain the tonal regions in the synthesis signal;

it is a block circuit diagram of a preferred embodiment based on the HFC high frequency contents;

shows a preferred implementation of the device of the invention with additional functionality to produce the direct sound channels L, R, C;

shows a stereo playback scenario;

shows a multichannel playback scenario in which all direct sound sources are reproduced through the front channels; Y

It shows a multichannel playback scenario in which sound sources can also be played back channels.

Figure 1 shows a device of the invention for generating an ambient signal 10 suitable for broadcasting through speakers for which a special speaker signal has not been transmitted. Speakers of this type are usually the rear speakers or the surround speakers, as shown by way of example in Figure 10 and Figure 11 in Ls, Rs.

The device shown in Figure 1 includes a transient detector 11 for detecting a transient period (shown in Figure 2 in 20) in which a test signal comprises a transient region. Although several implementations of the transient detector are described herein, it should be noted that any other method can be used to detect transients, such as, for example, those found in an MPEG-4 audio encoder, in which the switching Short to long windows are performed depending on a transient detection. In other fields of audio signal processing transient detectors are also used that

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

55

60

they are able to detect fast and strong variations of the envelope of a time signal. The orders of magnitude by way of example to be detected are variations of the envelope that in a period of 1 ms refer to variations of equal or more than 100% of the amplitude of the envelope.

The transient detector 11 is coupled to a synthesis signal generator 12 that is designed to generate a synthesis signal 13 that meets both conditions, that is, the transient condition on one side and the continuity condition on the other. The transitional condition is that the synthesis signal has a flatter time course than the test signal in the transitional region, while the continuity condition is that the intensity of the synthesis signal in the transitional region deviates from an intensity of an anterior or posterior part of the test signal less than a predetermined threshold. Preferably, the threshold is a relative threshold and is at a value = 2.5, in which values = 1.5 are even preferred. This means that the intensity of the signal in the transient region is at most 1.5 times or 0.66 times the intensity of a previous non-transient part or a subsequent non-transient part of the test signal. Therefore, it is guaranteed that a transient suppression does not result in a variation in amplitude and / or a disturbance intensity variation.

The threshold can also be made using a confidence interval of 80% or less, which is determined using historical values.

The intensity measurements that can be used for the present invention include the energy obtained by adding the sample squares or the squares of spectral value of a block, or a power measurement that can be obtained taking into account the length of the temporary block, or even a measurement that adds the magnitudes of the spectral values in a band in a weighted or unweighted way, in which this special measure that also represents an intensity refers to the high frequency contents when the band in which the sum takes place is the higher frequency band of the test signal or in general the higher frequencies will be weighted more strongly compared to the lower frequencies or have a stronger influence on the final result.

Next, the synthesis signal generator generates a synthesis signal used by a signal substitute 14 to use the synthesis signal instead of the corresponding region of the original test signal, to finally provide the ambient signal 10. Signal substitute 14 receives, apart from the synthesis signal through line 13, the test signal through a line 15, as indicated in Figure 1. The transient detector 11 receives the test signal through of an input line 16 and provides a transient information through an output line 17 to the synthesis signal generator 12 in order to generate the synthesis signal using the test signal provided through a line 18.

In the special embodiments of the present invention, a non-overlapping block process is used, as illustrated in Figure 2a, or an overlapping block process, as illustrated in Figure 2. In the non-overlapping block process of the Figure 2a, a test signal 21 is preferably divided into blocks of equal length having a special block length. Next, the transient detector detects a transient 22 in the transient period 20. Therefore, the transient 22 is in the transient period 20 of Figure 2a, the result being that the transient detector 11 provides an output signal to through its output line 17 which communicates to the synthesis signal generator 12 that the signal synthesis must start. While the blocks before and after the transitional period 20 directly represent the corresponding parts of the ambient signal 10 with the exception of cross-attenuation in a cross-attenuation region 23, the block of the test signal corresponding to the transitional period 20 is synthesized then by means of the synthesis signal generator and then it is used by signal substitute 14 instead of the original block of the test signal in the ambient signal.

As will be explained below, in the preferred embodiments the test signal block, which takes place in the frequency domain, is processed. This results in the synthesis signal in a block limit having a sample value that may differ considerably from a sample that is the last sample of the previous block in the test signal. In order to eliminate such block limit disturbances that may arise, it is preferred in the embodiment shown in Figure 2a to cross-mitigate a block before a transitional period to the synthesis signal in the transitional period, for example by adding the first sample of the synthesis signal generated to, for example, the last ten samples of the previous block that will be weighted according to the example cross-attenuation function according to the slow input function in Figure 3. At at the same time, the last sample of the previous block is added, according to the slow shutdown function of Figure 3, to the first samples or to the samples that follow the first sample, of the synthesized block that will be weighted according to the function of slow introduction in the transitional period to provide cross-attenuation. Correspondingly, the same method can be applied in the region of subsequent cross-attenuation, that is, when it is passed from the transitional post-block to the ambient signal block not influenced by the transients.

In order to further reduce block limit disturbances of this type, the processing of

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

55

60

overlay, as shown in figure 2b. In the embodiment shown in Figure 2b, the transient detector detects block regions represented by surrounded numbers (1), (2), (3), (4), (5), (6). A transient is detected at 22. The result is that compared to Figure 2a, there is a longer transition period 20 since the transient has been detected at position 22, both in block 4 and in block 5. Therefore , the synthesis signal generator 12 of Figure 1 will produce the synthesis signals for both block 4 and block 5. While for blocks prior to the three transitional region regions A, B, C, the signal of test has no transients and therefore is taken directly for the ambient signal, regions A, B, C are replaced by the signal substitute 14 of Figure 1 by the parts A, B, C produced by the signal generators of synthesis. A part A is produced by adding the second half of block 3 of the test signal not influenced by the transients to the first half of the synthesis signal produced by block 4. The second part B of the transition period 20 is provided by adding the second half of the synthesis signal produced by block 4 to the first half of the synthesis signal produced by block 5 and replaced by the signal substitute as a corresponding part of the ambient signal 10. The third part C of the transitional period 20 is produced by adding the second half of the block 5 produced by the synthesis signal generator to the first half of the block 6 that is no longer influenced by transients and written by the signal substitute 14 for the ambient signal.

The slow shutdown function shown in Figure 3 will be discussed in more detail below. Therefore, this slow shutdown function can be used to provide, when it comes to block processing with non-overlapping blocks, a smooth block transition from a non-synthesized block to a synthesized block and also to provide a smooth transition of a block synthesized back to a non-synthesized block. As an alternative, a cross-attenuation function can also be used to once again perform cross-attenuation to the original test signal, in particular, when a synthesis signal has been produced for a certain specific number of blocks. Since there is a probability that the synthesis signal, due to extrapolation, has deviated considerably from the test signal, a sharp turn back to the test signal in certain cases will result in audible disturbances. Therefore, it is preferred to perform the slow cross attenuation according to the slow input / slow shutdown function of Figure 3, producing, for a block in which no more transients have been detected, a synthesis signal consisting of a 90% of the last synthesized block and 10% of the current test block. In the following block, the ratio can be changed to 80%: 20% until, after a certain number of blocks, the synthesis signal is slowly turned off completely and the current test signal not affected by the transients is slowly introduced from completely new.

Subsequently, a preferred implementation of a part of the synthesis signal generator 12 will be treated with reference to Figure 4. For this, the time signal representing a block of the test signal is converted into a frequency domain representation or a sub-band representation by a converter 40 which may include a transformer or a bank of analysis filters. The spectral representation in the form of spectral coefficients or sub-band signals can then, as illustrated in 41, be replaced by the information in an extrapolated spectral representation and / or extrapolated sub-band signals if this is a block of the time signal in which a transient has been detected. Subsequently, the spectral representation is fed, perhaps using additional information due to an extrapolation, to a softener 42 that influences the spectral values such that the time course of the underlying signal is softened. In the case of a filter bank, this softener 42 will influence the sub-band signals in such a way that the time course of the underlying signal of the sub-band signals is smoother than before smoothing. Then, in block 43, an inverse conversion to the time domain is performed, in which either a retransformer or a bank of synthesis filters is used to finally reach a time signal 44 that has a smoother course than The time signal at the entrance of stage 40, however, has a considerable amount of energy not influenced by smoothing. In addition, smoothing has been performed in such a way that the energy of the smoothed time signal 44 does not differ from the energy of the previous time signal by more than the threshold.

Therefore, in the present invention, a total energy manipulation of the time signal energy can take place. However, only the transients are attenuated, while the tonal parts continue and / or are synthesized from the story by synthesizing the signal in the transitional period by a prediction that uses a non-transient signal from the past.

If, however, the energy is not touched, as when in a randomization or in a spectral prediction, the smoothing has resulted in the energy being distributed more evenly over the block in such a way that a time course has been generated softer without changing, however, considerably the energy of the sample block of the test signal. This is sufficient in most cases and it is guaranteed that the user will hear a test signal always complying with the continuity condition. Only if the transient results in a considerable increase in energy, taking into account the entire block, will it be only smoothing, that is, a more uniform distribution of energy over the block, which is no longer sufficient, and can the clipping be performed. of controlled signal.

Well-known methods, which include avoiding the location of direct sound sources in the rear channels, are delaying the rear channels for a few milliseconds. This solution does not result in the suppression of transients, but rather tries to "mask" the transients using the precedence effect. The effect

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

55

60

of precedence is that the ofdo supposes that a sound source is where some of this sound source is heard for the first time, in which what is heard next, from this sound source may very well be louder or Come from a different address. However, this solution is a disadvantage because very short sound events that have pronounced transients can often still be offered and then they are perceived twice, by a front speaker and a few milliseconds later by the rear channels, causing an impression of unpleasant hearing.

Commercially available matrix decoders, such as, for example, Dolby Pro Logic II or Logic 7, have the ability to mix up unprocessed 2-channel stereo files in multichannel surround sound files even though they are not directly designed For this task. These matrix decoders are often not able to suppress transient tones in the rear channels, resulting in a signal that does not meet the requirements for freedom and transient continuity in amplitude and / or intensity.

However, channel regions where there are transients according to the invention are detected and attenuated. However, simply attenuating the complete signal in these periods will result in a modulation of the amplitude of the ambient signal and will be perceived as unpleasant or even as a disturbance. Therefore, this could avoid the sensation of quality of the signal of an extradited or processed environment. To overcome this unpleasant amplitude modulation effect, a transient suppression occurs according to the invention without preventing the continuity of the synthesis signal and / or the ambient signal. In this document, an input signal, such as, for example, an upmix signal, to the extent that is achieved by an ascending matrix mixer, is used for the rear channels or a signal having similar characteristics and a Similar field of application is analyzed to detect if there is a transient.

If a transient is detected, the block currently processed will be replaced by a replacement signal that has a flat (non-transient) temporary envelope. This substitution signal is produced either by the previous signal parts, where there have been no transients or is produced by the block currently processed by a processing stage that performs the temporary envelope and / or a thinner structure of the flattest signal , or produced by a combination of both methods.

The substitution signal produced by the previous parts is produced, for example, by an extrapolation of the previous energy levels of the signal or by copying / repeating the previous signal parts with no transient region of the signal.

The "flattening" of the temporal fine structure or the fine time signal on the basis of the block processed herein can, for example, be performed in a manner illustrated below with reference to Figures 5a, 5b or 5c.

The absolute values of the spectral coefficients can be randomized within a limited region that extends around the extrapolated spectral coefficients or magnitudes thereof, as will be explained later in relation to Figure 5c.

Alternatively or additionally, the phases and / or signs of the spectral coefficients of the processing block in which the transient can be randomized by a randomizer 50. For this, a short-term spectrum of the block of the test signal considered is produced and They calculate the complex spectral values obtained according to the magnitude and phase to then randomize the phases of the spectral values. If a transformation is used, which can only solve the +/- 180 ° phases, that is, it can only provide spectral values with a positive and negative sign, the signals can also be randomized to obtain a short-term spectrum that has the randomized phases / signs of a flatter time course of the corresponding time signal.

This approach is based on the fact that a rapid change in a time signal will only be possible if the phases of the fundamental wave that underlie this transitional region and the respective harmonics are in a special relationship. If a randomization of the phases is achieved, this will result in softening the transient region since the special interaction of the individual sinusoidal oscillation phases mapped by the spectral values is no longer there.

An alternative implementation is illustrated in Figure 5b that uses a predictor 51 that is implemented to make a short-term spectrum prediction along the frequency. Such a predictor is illustrated in J. Herre, J.D Johnston; “Exploiting Both Time and Frequency Structure in a System that Uses an Analysis / Synthesis Filterbank with High Frequency Resolution,” 103rd AES Convention, New York 1997, Preprint 4519.

Once again, a short-term spectrum is produced that has a transitory course in its associated time signal. Normally, using an open loop predictor, a current spectral value of the spectrum is predicted in the short term by means of a previous spectral value or a plurality of previous spectral values, in which the predicted spectral value could then be subtracted from the value real spectral to obtain a spectral residual value.

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

55

60

While the spectral residual value of a typical frequency prediction represents that value that is of interest and carries information along with the coefficients of a prediction filter, a certain prediction filter is inventively preset and the spectral values of the spectrum in the short term they are replaced by the predicted spectral values using this prediction filter, while the prediction error signal is no longer used.

However, the actual spectral values of defective prediction obtained below have a flatter time course than the original short-term spectrum, but still have approximately the same amount of energy in such a way that both the transient condition and the condition of continuity, as illustrated in relation to the synthesis signal generator 12 of Figure 1. A simple preferred implementation of the prediction filter is simply to use a value of a spectral line having a smaller index as a prediction value for a Current spectral line.

In general, the extrapolated signal can cross-mitigate with the original signal after a certain time, instead of abruptly changing to avoid long-term extrapolation disturbances.

In addition, it is preferred, as illustrated with reference to Figure 6, to detect the tonal parts / bands by a detector 60 and not influence them by the synthesis signal generator, but combine them in a mixer / combiner 61 with the synthesis signals for the transient bands to obtain, after the transformation or conversion in the time domain, which can take place in block 61, a time signal that has a flatter time course, which, without However, it still includes tonal bands, that is, parts that have not been transient, unaltered.

Therefore, stationary / tonal frequency components are detected in the input signal that have been present, for example, during the duration of the transient only in parts of the spectrum and a replacement signal is generated that includes an extrapolation of the components of past stationary / tonal signals and the stationary / tonal frequency components in the current block.

Subsequently, an implementation of the present invention using an implicit and no longer explicit transient detector will be illustrated with reference to Figure 5c. The means 53 for calculating the intensity of a block and a previous block is shown in Figure 5c. A measure of the intensity of a block of processed signal is, for example, energy or high frequency content (HFC) or another measure that is based on the spectral values, time samples, energy, power or other measurement of the signal related to amplitude. Next, it is determined by means 54 if an intensity increases from one block to the next beyond a threshold. If this is the case, the spectral values of the processed block are limited in such a way that their intensities do not exceed the intensity of the previous signal block by more than a certain relative or absolute threshold such that at least the overall domain of the transients. This limitation is made in the medium 55 which is implemented to limit, if a demand for a limitation has been detected, that is, implicitly detecting a transient, the spectral values either individually or globally. An individual limitation should calculate an increase in energy for the spectral values or for the bands and the spectral values and / or the energy bands that increase only to a maximum increase in energy and that is cut further.

The means 55 for limiting the spectral values therefore limits the spectral values individually or globally, in which an individual limitation is that only the spectral values that increase beyond a threshold are limited and preferably limited to this threshold, while the other spectral values that do not increase so strongly are not influenced. As an alternative, however, it will be more favorable in some cases and easier with respect to the calculation of complexity to limit all spectral values to the same absolute or relative measure if two strong increases have been determined.

In addition, it is preferred to perform post-processing of the spectral values limited by means 56 for post-processing, in which this post-treatment may be a randomization, as described in Figure 5a, or a prediction, as described in figure 5b. The order of processing by means 55 and 66 can also be reversed in such a way that in the first randomization and / or prediction process it is carried out with a block for which a transient has been detected, in which only then a limitation is made intensity according to the processing in block 55.

With respect to Figure 5c, it should be noted that the t / f block represents a time / frequency domain conversion 57, in which a conversion from the time domain to the frequency domain can also be filtered by means of a bank of Analysis filter in such a way that in this case the spectral representation consists of sub-band signals and not individual components.

Subsequently, an especially preferred embodiment of the present invention will be discussed with reference to Figure 7. The transient detector, as shown in Figures 1 to 11, in this embodiment includes a means 71 for calculating the high frequency contents (HFC ) for each block downstream of the medium to calculate the HFC 72 in the long term. Next, a comparator 73 will detect if there is a transitory or if there is a transitional period in

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

55

60

the one that exists a transitory. In particular, the medium 71 is implemented to calculate the weighted high frequency contents (HFC) for each block of the original left signal and the original right signal. Alternatively, an HFC can be calculated for each channel. The HFC is the weighted sum of the absolute values of all frequency lines in a block, increasing the weighting factors from the lowest to the highest frequencies. The HFC is calculated as follows:

HFC = sum (X (f) w (f)),

in which X (f) are the spectral coefficients for certain frequencies, w (f) being weighting factors for certain frequencies.

Due to the fact that the weighting factors increase from the lowest to the highest frequencies, it is guaranteed that in the value of the HFC, the energy in the higher frequency components is weighted in comparison to the energy in the frequency components more low. An energy in the higher spectral components is better than a index for a transient than an energy in the lower spectral components. In the implementation, all spectral components can be used to calculate the HFC. As an alternative, the calculation of the HFC can also be made from a threshold value that is approximately in the central region of the spectrum such that the lower spectral coefficients do not play a role in the calculation of the HFC.

In addition, a long-term average value of HFC is also calculated as HFC 'in at least three and preferably in five previous blocks. If it is determined in the middle 73 that the HFC in the current block deviates from the long-term average value of the HFC 'a factor greater than a constant factor c, a number> 1.0 that is used as the constant factor c, will detect a transient. The threshold depends on the type of the floating average value. If the average floating value is an average value in which the story becomes stronger compared to the most current block, that is, a slower average value, the threshold will be closer to 1 than in the case where the history introduces the average floating value to a lesser extent. In this case, the threshold is farther than 1.

If a transient is detected, as indicated for the medium 74 to calculate the average value by means 73, the average value of the past absolute values of each frequency line (the spectral coefficients) over a range of defined time, such as, for example, five blocks. In addition, an Amax prediction reliability interval is calculated for the extrapolated absolute values. Absolute extrapolated values go randomly within this Amax range. In order to achieve this, a calculation is made according to an equation as shown in Figure 7 in the middle 75. RN represents a random number, Amax represents the reliability range, SW is a spectral value, as calculated by means of calculation 75, and SWm is the spectral value that results as an average value of several previous blocks, as calculated by block 74. Means 75 is therefore implemented to evaluate the following equation:

SW = SWm + RN • Amax

In order to avoid the repetition effects that may arise when a detected transient is too long, cross attenuation is performed on the values extrapolated with the original values, at a time when a fixed time interval has passed, for example, three Synthesis signal blocks that have to be present from which the original signal must arrive again. However, if the transitional period is shorter than the three blocks, it will be preferred not to perform cross-attenuation, since it can be assumed that the extrapolated signals have not yet deviated too much from the original signals. Cross-attenuation can take place either before or after a conversion to the time domain or, preferably, after a conversion to the time domain, as illustrated in Figure 7 in 76, to obtain the synthesis signal.

In one implementation, the concept of the invention can be integrated into a process of extracting an environment signal or used as a separate post-processing stage using an existing environment signal that, however, still includes unwanted transients before The processing of the invention.

The process steps of the invention can be performed in the frequency domain by frequency line or in the subbands. However, they can also be performed only in part in the frequency domain normally above a certain frequency limit or in a time domain exclusively or in a combination of time and frequency domains.

Figure 8 shows a preferred embodiment of the present invention in which the device for generating an ambient signal is not only implemented to generate the ambient signals for an output 80 for a left ambient channel and an output 81 for a communication channel. right environment. In addition, the device of the invention includes an ascending mixer 82 for generating signals for the left channel L, the right channel R, the central channel C and preferably also for the LFE channel as shown in Figure 8. Both the combination of the

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

55

60

transient detector 12, the synthesis generator 14 and a signal substitute 16 and the ascending mixer 82 are fed by a decoder 84. The decoder 84 is implemented to receive and process a bit stream 85 to provide a mono signal or a signal stereo 86 on the output side. The bit stream can be an MP3 bit stream or an MP3 file or it can be an AAC file or it can be a representation of a parameterically encoded multichannel signal. Therefore, bit stream 85 can be, for example, a representation parameter of the left channel, the right channel and the central channel, in which a transmission channel and several signals for the second and third channels are contained, This processing is known from the processing of multiple BCC channels. Then, the decoder 84 signals a BCC decoder that not only provides a mono or stereo signal, but also provides a three-channel signal, which, however, does not include data in the two surround channels Ls, Rs. In one implementation, the test signal will in this case be a mono signal, a stereo signal or even a multichannel signal which, however, does not include special speaker signals for the surround sound channels Ls, Rs.

It should be noted that either the same ambient signal can be calculated for both surround channels or a special signal for each surround channel. In the first case, the test signal and / or the surround sound signal are obtained, for example, from a sum of the left and right channels. In the other case, the ambient signal for the left surround sound channel is calculated, for example, from the left channel and the ambient signal for the right channel is calculated from the right channel.

One aspect of the present invention relates to a device for generating a suitable ambient signal to be output through loudspeakers Ls, Rs for which there is no suitable speaker signal, comprising: a transient detector 11 for detecting a transitional period 20 in which a test signal comprises a transitional region 22; a synthesis signal generator 12 to generate a synthesis signal during the transitional period 20, the synthesis signal generator 12 being implemented to generate a synthesis signal comprising a flatter time course than the test signal in the transitional period 20 and whose intensity deviates from an intensity of an anterior or posterior part of the test signal less than a predetermined threshold; and a substitute for signal 14 to replace the test signal in the transitional period with the synthesis signal to obtain the ambient signal.

In one embodiment, the device is implemented for block processing to process subsequent blocks of discrete time samples in an overlapping or non-overlapping manner.

In one embodiment, the transient detector 11 is implemented to calculate the intensity values for the subsequent blocks and to detect a transient period 20 when an intensity value of a block differs from a previous or subsequent intensity value more than a transient threshold predetermined.

In one embodiment, the synthesis signal generator 12 is implemented to limit, for a block in the transitional period 20, a plurality of spectral values that represent a short-term spectrum of the block such that its intensities differ from the intensity of a previous or subsequent block or transient less than the predetermined threshold.

In one embodiment, the synthesis signal generator 12 is implemented to randomize the complex spectral values that represent a short-term spectrum of the block including the transitional period 20 with respect to its phases or signs.

In one embodiment, the synthesis signal generator 12 is implemented to perform the prediction processing 51 on the frequency to obtain a prediction spectrum of the associated time signal of which it comprises a flatter time course than an associated time signal. to a spectrum before the prediction process along the frequency.

According to the invention, the transient detector 11 is implemented to calculate 61 the high frequency contents for a block of the test signal; wherein the transient detector 11 is implemented to compare 73 the weighted HF contents for a floating average value with a plurality of anterior or posterior blocks without transients, in which the transient detector 11 is implemented to detect a transient of a block when the Hf contents of a current block exceeds the floating average value by more than a threshold c.

In one embodiment, the transient detector is implemented to use a threshold that is selected based on the type of calculation of the average floating value and that is closer to it when history has a strong influence on the average floating value, and is more away from it when history has a comparatively smaller influence on the average floating value.

In accordance with the invention, the synthesis signal generator is implemented to calculate 74, for each spectral value of a short-term spectrum of a plurality of blocks, an average value using the corresponding spectral values of the plurality of blocks to obtain a average value spectrum, to calculate, for

5

10

fifteen

twenty

25

30

35

40

Four. Five

fifty

55

60

spectral values, the deviations that differ from the spectral values and that are smaller than a maximum Amax deviation, and to add the mean deviations and values to the spectral values to obtain a processed spectrum.

In one embodiment, the synthesis signal generator 12 is implemented to calculate the synthesis signal from the signal parts of the test signals before or after the transitional period, from the test signal in the transitional period after of smoothing the time course of it or from a combination of the signal parts of the test signal and the test signal after smoothing.

In one embodiment, the synthesis signal generator 12 is implemented to copy the signal parts of the test signal before or after the transitional period.

In one embodiment, the synthesis signal generator 12 is implemented to randomize, in a predetermined domain, the extrapolated spectral values obtained from the test signal outside the transitional period.

In one embodiment, the synthesis signal generator 12 is implemented to mix, when the transitional period has a longer duration than a predetermined time, for times after the predetermined period, the synthesis signal values with the signal values of The test signal.

In one embodiment, the signal substitute 14 is implemented to perform cross-attenuation from one part before the transitional period to the transitional period according to a cross-function or cross-attenuation from the transitional period to a part after the transitional period of agreement with a cross-attenuation function.

In one embodiment, the synthesis signal generator 12 is implemented to calculate 40, 41, 42 a short-term spectrum of the synthesis signal with spectral values, to convert the short-term spectrum into a temporal representation that represents the Synthesis signal 44.

In one embodiment, the synthesis signal generator 12 is implemented to calculate 40, 41, 42 a short-term spectrum of the synthesis signal with sub-band signals, and to convert the short-term spectrum with signals from sub-band in a temporal representation representing 43 the synthesis signal.

In one embodiment, the synthesis signal generator 12 is implemented to generate the synthesis signal such that the predetermined threshold is less than or equal to a factor of 2.

In one embodiment, the synthesis signal generator 12 is implemented to use a preset selective band threshold or a single threshold for the entire spectrum.

In one embodiment, the device further comprises means for processing a left channel signal and a right channel signal to extract the test signal.

In one embodiment, the device comprises a two to three mixer 82 to generate a left channel, a right channel and a central channel from a transmitted stereo or mono signal; in which the synthesis signal generator 12 is implemented to provide the same ambient signal to the rear left and right rear channels or to scale the test signal so that the rear left channel and the rear right channel can receive different versions climbing the ambient signal or to calculate two special ambient signals for two surround channels.

In addition, the present invention relates to a method for generating a suitable ambient signal to be output through loudspeakers Ls, Rs for which there is no suitable loudspeaker signal, comprising the steps of: detecting a transient period 20 wherein a test signal comprises a transient region 22; generating 12 a synthesis signal during the transitional period 20, implementing the synthesis signal generator 12 to generate a synthesis signal that comprises a flatter time course than the test signal in the transitional period 20 and whose intensity deviates from an intensity of an anterior or posterior part of the test signal less than a predetermined threshold; and replace 14 the test signal in the transitional period 20 with the synthesis signal to obtain the ambient signal.

Depending on the circumstances, the method of the invention can be implemented either in hardware or in software. The implementation can be in a digital storage medium, in particular, in a disk or CD that has control signals that can be read electronically, which can be operated together with a programmable computer system in such a way that the method is executed. In general, the invention is therefore also in a computer program product that has a program code stored on a machine-readable medium to perform the method of the invention when the computer program product is run on a computer. Said of another

Thus, the invention can therefore also be carried out as a computer program that has a program code to perform the method when the computer program is run on a computer.

Claims (10)

5 10 fifteen twenty 25 30 35 40 Four. Five fifty 55 60 1. A device for generating a suitable ambient signal to be output through speakers (Ls, Rs) for which there is no suitable speaker signal, comprising: a transient detector (11) for detecting a transient (22) in a block of a test signal to obtain a transient period (20) comprising the block of the test signal; a synthesis signal generator (12) to generate a synthesis signal during the transitional period (20), the synthesis signal generator (12) being implemented to generate a synthesis signal comprising a flatter course of time than the signal of test in the transitional period (20) and whose intensity deviates from an intensity of a part of the test signal before the transitional period (20) or after the transitional period (20) by less than a predetermined threshold; Y a signal substitute (14) to replace the test signal in the transitional period with the synthesis signal to obtain the ambient signal, wherein the transient detector (11) is implemented to calculate the weighted high frequency contents for the test signal block (71), in which the weighted high frequency contents are a weighted sum of the absolute values of all frequency lines in the block with increasing weighting factors from the lowest to the highest frequencies; wherein the transient detector (11) is implemented to compare (73) the weighted high frequency contents of the block for a floating average value (72) with the weighted high frequency contents of a plurality of previous or subsequent blocks of the block without any transitory, wherein the transient detector (11) is implemented to detect a transient for the block when the weighted high frequency contents of the block exceed the floating average value in more than one threshold (c), in which the signal generator of Synthesis is implemented to calculate (74), when the synthesis signal is generated for all spectral values of a short-term spectrum of a plurality of blocks, which form the part of the test signal before the transitional period (20) or after the transitional period (20), an average value that uses the corresponding spectral values of the plurality of blocks to obtain an average value spectrum for the block, calculate, for the spectral values of the average value spectrum for the block, the deviations that differ from the spectral values of the average value spectrum for the block and that are smaller than a maximum deviation (Amax), and add (75) the deviations and the mean value of the spectral values to obtain the spectral values for the block, and convert (43, 76) the spectral values for the block into a temporal representation that represents the synthesis signal (44). 2. The device according to claim 1, which is implemented to process the block in order to divide the test signal into successive and overlapping or non-overlapping blocks of discrete time sample values. 3. The device according to claim 1, wherein the threshold (c) is selected based on the type of calculation of the average floating value and that is closer to one when history has a stronger influence on the average value floating, and is further from one when history has a comparatively smaller influence on the average floating value. The device according to one of the preceding claims, wherein the synthesis signal generator (12) is implemented to calculate the synthesis signal from signal parts of the test signals before or after the transitional period. , from the test signal in the transitional period after smoothing the time course thereof or from a combination of the signal parts of the test signal and the test signal after smoothing. The device according to one of the preceding claims, wherein the synthesis signal generator (12) is implemented to generate the synthesis signal such that the predetermined threshold is less than or equal to a factor of 2 . The device according to one of the preceding claims, wherein the predetermined threshold is a preset selective band threshold or a single threshold for the entire spectrum. 7. The device according to one of the preceding claims, further comprising: Extraction means for processing a left channel signal and a right channel signal to extract the test signal. 8. The device according to one of the preceding claims, further comprising: 5 10 fifteen twenty 25 30 35 40 Four. Five fifty a two to three mixer (82) to generate a left channel, a right channel and a central channel from a transmitted stereo signal; Y wherein the synthesis signal generator (12) is implemented to provide the ambient signal to a rear left channel (Ls) and the ambient signal to a rear right channel (Rs), in which the left rear channel and the right rear channel is the channels for which there is no suitable speaker signal, or in which the synthesis signal generator (12) is implemented to scale the test signal in such a way that versions of the ambient signal are obtained climbing differently for the left rear channel and the right rear channel, or wherein the synthesis signal generator (12) is implemented to calculate a first ambient signal for the rear left channel and to calculate a second ambient signal that differs from the first ambient signal for the rear right channel. 9. A method for generating a suitable ambient signal to be output through speakers (Ls, Rs) for which there is no suitable speaker signal, comprising the steps of: detecting (11) a transient (22) in a block of a test signal to obtain a transient period (20) comprising the block of the test signal; generate (12) a synthesis signal during the transitional period (20), implementing the synthesis signal generator (12) to generate a synthesis signal comprising a flatter time course than that of the test signal in the transitional period (20) and whose intensity deviates from an intensity of a part of the test signal before the transitional period (20) or after the transitional period (20) at less than a predetermined threshold; Y replace (14) the test signal in the transitional period (20) with the synthesis signal to obtain the ambient signal, in which at the detection stage (11) the weighted high frequency contents for the test signal block (71) are calculated, in which the weighted high frequency contents are a weighted sum of the absolute values of all the frequency lines in the block with increasing weighting factors from the lowest to the highest frequencies; wherein at the stage of detecting (11) the weighted high frequency contents of the block are compared (73) with a floating average value (72) along the weighted high frequency contents of a plurality of previous or subsequent blocks of the block without any transitory, in which at the stage of detecting (11) a transient is detected for the block when the weighted high frequency contents of the block exceed the floating average value by more than one threshold (c), in which at the stage of generating (12) when the synthesis signal is generated for all the spectral values of a short-term spectrum of a plurality of blocks, which form the part of the test signal prior to the transitional period (20 ) or after the transitional period (20), an average value (74) is calculated using the corresponding spectral values of the plurality of blocks to obtain an average value spectrum for the block, in which at the stage of generating (12) for the spectral values of the average value spectrum for the block, deviations are calculated that differ from the spectral values of the average value spectrum for the block and that are smaller than a maximum deviation (Amax), and in which at the stage of generating (12) the deviations and the average values, the spectral values are added (75) to obtain the spectral values for the block, and in which at the stage of generating (12), the spectral values for the block are converted (43, 76) into a temporal representation that represents the synthesis signal. 10. A computer program for performing a method according to claim 9, when the method is executed on a computer.